基于数学变换和连续小波变换的苹果园土壤总氮含量高光谱反演

doi:10.14088/j.cnki.issn0439-8114.2025.11.022

摘要/Abstract

摘要： 为快速测定苹果园土壤总氮(TN)含量,实现精准施肥,以山东省果树研究所天平湖基地苹果园的土壤为研究对象,测定其光谱反射率（R）并进行数学变换（倒数、对数、平方根及一阶微分等）、连续小波变换(Continuous wavelet transform,CWT)及数学变换结合CWT处理,采用Pearson相关分析法进行特征提取,基于支持向量回归(Support vector regression,SVR)构建土壤TN含量的高光谱反演模型。结果表明,原始光谱数据经CWT处理后,在2¹~2¹⁰尺度内,小波系数与土壤TN含量的相关性呈先升高后降低的趋势,中等尺度能有效抑制噪声干扰并增强光谱与TN的相关性,效果显著。数学变换、CWT均能有效挖掘光谱的细节特征,CWT的效果总体上优于数学变换。数学变换结合CWT能显著提升模型反演精度,其中,1/R-CWT-2⁸模型的反演效果最优(R²=0.73,RMSE=0.12 g/kg,RPD= 1.85)。该模型预测的土壤TN含量与实测值的拟合曲线更接近1:1线,模型预测精度较高。综上,高光谱技术可以作为苹果园土壤TN含量的无损检测手段,数学变换结合CWT构建的高光谱反演模型可以更加精准地预测土壤TN含量。

关键词: 数学变换, 连续小波变换（CWT）, 苹果园, 土壤, 总氮含量, 高光谱, 反演

Abstract: To rapidly determine the soil total nitrogen (TN) content in an apple orchard and achieve precise fertilization, soil from the apple orchard at the Tianping Lake Base of the Shandong Institute of Pomology was taken as the research object. Its spectral reflectance (R) was measured and subjected to mathematical transformation (reciprocal, logarithm, square root, first-order derivative, etc.), continuous wavelet transform (CWT), and combined mathematical transformation and CWT processing. The Pearson correlation analysis method was used for feature extraction, and a hyperspectral inversion model for soil TN content was constructed based on support vector regression (SVR). The results showed that after processing the original spectral data with CWT, within the scale range of 2¹ to 2¹⁰, the correlation between the wavelet coefficients and soil TN content first increased and then decreased. Medium scales effectively suppressed noise interference and enhanced the correlation between the spectrum and TN, with a significant effect. Both mathematical transformation and CWT effectively mined the detailed features of the spectrum, with the effect of CWT generally being superior to that of mathematical transformation. The combination of mathematical transformation and CWT significantly improved the model inversion accuracy. Among them, the 1/R-CWT-2⁸ model demonstrated the best performance (R²=0.73, RMSE=0.12 g/kg, RPD = 1.85). The fitting curve between the soil TN content predicted by this model and the measured values was closer to the 1:1 line, indicating high model prediction accuracy. In conclusion, hyperspectral technology could be used as a non-destructive testing method for soil TN content in apple orchards. The hyperspectral inversion model constructed by combining mathematical transformation and CWT could more accurately predict soil TN content.

Key words: mathematical transformation, continuous wavelet transform (CWT), apple orchard, soil, total nitrogen content, hyperspectral, inversion

中图分类号:

S153
S661.1

熊超群, 荀咪, 李健, 安淼, 李国田. 基于数学变换和连续小波变换的苹果园土壤总氮含量高光谱反演[J]. 湖北农业科学, 2025, 64(11): 165-170.

XIONG Chao-qun, XUN Mi, LI Jian, AN Miao, LI Guo-tian. Hyperspectral inversion of soil total nitrogen content in an apple orchard based on mathematical transformation and continuous wavelet transform[J]. HUBEI AGRICULTURAL SCIENCES, 2025, 64(11): 165-170.

参考文献

[1] 陈印,刚成诚,刘欢欢,等.基于GEE的苹果园地遥感信息提取研究——以陕西省渭北旱塬区为例[J].西北林学院学报,2024,39(1):36-43.
[2] 高义民,同延安,路永莉,等.陕西渭北红富士苹果园土壤有效养分及长期施肥对产量的影响[J].园艺学报,2013,40(4):613-622.
[3] 曾鹏宗,王旺,袁敏鑫,等.基于无人机遥感的苹果树冠层氮含量反演研究[J].山东农业科学,2024,56(10):167-173.
[4] 王梦樊,吉艳芝,刘娉妤,等.果园土壤氮损失监测方法研究进展[J].植物营养与肥料学报,2023,29(3):553-572.
[5] 郝琨. 黄土高原涌泉根灌土壤水氮耦合特性及对山地苹果的协同效应研究[D].西安:西安理工大学,2022.
[6] HORTA A, AZEVEDO L, NEVES J, et al.Integrating portable X-ray fluorescence (pXRF) measurement uncertainty for accurate soil contamination mapping[J]. Geoderma, 2021, 382: 114712.
[7] SUN W C, LIU S, ZHANG X, et al.Estimation of soil organic matter content using selected spectral subset of hyperspectral data[J]. Geoderma, 2022, 409: 115653.
[8] MALLAH NOWKANDEH S, NOROOZI A A, HOMAEE M.Estimating soil organic matter content from Hyperion reflectance images using PLSR, PCR, MinR and SWR models in semi-arid regions of Iran[J]. Environmental development, 2018, 25: 23-32.
[9] 叶淼,朱琳,刘旭东,等.基于连续小波变换、SHAP和XGBoost的土壤有机质含量高光谱反演[J].环境科学,2024,45(4):2280-2291.
[10] 王海江,刘凡,YUNGER J A,等.不同粒径处理的土壤全氮含量高光谱特征拟合模型[J].农业机械学报,2019,50(2):195-204.
[11] ABULAITI Y, SAWUT M, MAIMAITIAILI B, et al.A possible fractional order derivative and optimized spectral indices for assessing total nitrogen content in cotton[J]. Computers and electronics in agriculture, 2020, 171: 105275.
[12] 龚明冲,汪泓,张磊,等.基于光谱指数与连续小波变换的土壤有机碳含量估算[J].激光与光电子学进展,2025,62(3):360-368.
[13] 于雷,洪永胜,周勇,等.连续小波变换高光谱数据的土壤有机质含量反演模型构建[J].光谱学与光谱分析,2016,36(5):1428-1433.
[14] 玉米提·买明,王雪梅.连续小波变换的土壤有机质含量高光谱估测[J].光谱学与光谱分析,2022,42(4):1278-1284.
[15] 王莉雯,卫亚星.湿地土壤全氮和全磷含量高光谱模型研究[J].生态学报,2016,36(16):5116-5125.
[16] 李嘉琦,冯宇华,陈署晃,等.基于高光谱的土壤有机质及全氮估测[J].新疆农业科学,2024,61(10):2491-2499.
[17] FARGE M.Wavelet transforms and their applications to turbulence[J]. Annual review of fluid mechanics, 1992, 24: 395-458.
[18] ZHANG S W,SHEN Q,NIE C J,et al.Hyperspectral inversion of heavy metal content in reclaimed soil from a mining wasteland based on different spectral transformation and modeling methods[J]. Spectrochimica acta part A: Molecular and biomolecular spectroscopy, 2019, 211: 393-400.
[19] 陈斌,贾建,张慧玉,等.基于脊波与小波变换的联合图像去噪方法[J].西北大学学报(自然科学版),2008,38(5):739-743.
[20] 国佳欣,朱青,赵小敏,等.不同土地利用类型下土壤有机碳含量的高光谱反演[J].应用生态学报,2020,31(3):863-871.
[21] WRIGHT A F, BAILEY J S.Organic carbon, total carbon, and total nitrogen determinations in soils of variable calcium carbonate contents using a Leco CN-2000 dry combustion analyzer[J]. Communications in soil science and plant analysis,2001, 32(19-20): 3243-3258.
[22] 管铖,刘明月,满卫东,等.连续小波结合随机森林算法估算互花米草叶片叶绿素含量[J].光谱学与光谱分析,2024,44(10):2993-3000.
[23] ZHAO L Y, TAN K, WANG X, et al.Hyperspectral feature selection for SOM prediction using deep reinforcement learning and multiple subset evaluation strategies[J]. Remote sensing, 2023, 15(1): 127.
[24] 王英杰,高欣梅,包淑梅,等.基于高光谱数据的盐渍化土壤含盐量预测模型研究[J].农业工程,2024,14(10):113-120.
[25] 郭艳萍,王雪梅,赵枫,等.基于最优数学和小波变换的绿洲耕层土壤盐分RF高光谱反演[J].农业工程学报,2025,41(3):83-93.
[26] 王延仓,金永涛,王晓宁,等.传统光谱变换与连续小波耦合定量反演潮土有机质含量[J].光谱学与光谱分析,2018,38(8):2571-2577.
[27] 安柏耸,王雪梅,黄晓宇,等.基于连续小波变换的土壤重金属镉含量的高光谱估测[J].地球与环境,2023,51(2):246-253.